The mercury model

Gridded emission data has been produced on five year intervals from 1990. The pre 2005 datasets were distributed based using a population proxy only. For the 2005 gridded emission dataset a new multi-layer model was developed. This model uses proxies like industrial CO₂ emissions and urban population which are better suited to serve as a surrogate for mercury emission.

The main sources of mercury emissions are the use of coal and artisanal gold mining. Lesser, but still significant contributors are for instance production of metals and cement, waste incineration and chlor-alkali processes. A spatial distribution model is being developed to model mercury emission to the atmosphere. The model domain is a 0.5x0.5° global raster. The model output includes total Hg, particulate Hg, Hg²⁺ and Hg0. The model also takes into account chimney/stack height (currently three classes; 0-50m, 50-150m, >150m).

The 2010 geospatial distribution model employed is essentially the same as was used for the 2005 inventory – i.e. a distribution of the national emissions estimates to point and area sources – but with a number of improvements. One major improvement is that the number of point sources to which emissions are allocated is significantly increased, especially for countries other than Europe and North America where few or no point sources were applied in the 2005 inventory distribution. Some of these point sources have associated mercury emissions estimates (e.g. as reported in national pollution release inventories). For point sources with no reported emission, estimates were made based on assumptions utilising information such as plant capacity, and fuel use, etc. and scaling of national total emissions for the sectors concerned. These methods will be more fully described in papers that are under preparation.

As a result of these improvements the proportion of the emissions that are designated as ‘area emissions’ that is emissions from diffuse sources or emissions from point sources that are not identified is far lower. This results in emissions being more ‘focussed’ to individual cells as is clearly shown on the emission map.